Repairing Poor Quality Well Log Data with Interactive Petrophysics using Auto Edit

Well log data is a key data source for subsurface analysis and petrophysical machine learning models. But, when that well log data is of poor quality, which could be caused by sensor issues, borehole conditions, or simple human error, it can have knock-on effects further down the interpretation pipeline.

Any errors in derived curves and properties can multiply as they go through the various interpretation stages. Small errors early on can become major problems later. This can lead to important decisions made on incorrect information.

For example, a poor quality bulk density log can result in miscalculated porosity estimates, which in turn can skew water saturation calculations, and feeds misleading inputs into geological models and simulations.

When we are working with good quality data, the value of the insights based on solid interpretation results can increase significantly. This in turn means that we have more reliable subsurface models that underpin investment decisions, future drilling plans and reservoir lifecycle management.

Fortunately, poor quality well log data does not have to derail the entire workflow. IP offers multiple tools to help quality check and repair your data before it causes trouble.

Auto Edit: Your Petrophysical Data Repair Kit

Curve Auto Edit is an Interactive Petrophysics (IP) tool which is used to simplify basic log editing using multiple log curve linear regressions. It has a systematic workflow that will automatically generate several regression sets based on zoning and supplied flags to indicate areas of bad data.

After the regressions have been generated the module identifies the best regression set and patches in the corrected data wherever the original data has been flagged as erroneous.

Key features:

• Flag Driven Logic: Use flags derived from the LogQC module or define your own through interactive crossplots to isolate problematic data, giving you full control over what data gets repaired.
• Flexible Input Handling: Auto Edit requires a Gamma Ray curve and a Resistivity curve as baseline inputs. Beyond that, you can provide additional log curves and Auto Edit will adapt to what is available and adjusts the regressions accordingly.
• Iterative Workflow: You can easily adjust flags, modify zone boundaries and quickly re-run regressions to refine your model without starting from scratch.
• Regression Based Patching: Auto Edit uses fast and efficient multiple regression predictions and sophisticated logic to systematically repair and reconstruct multiple well log measurements. The algorithm dynamically selects the best regression
• Smooth Transitions: You can apply smoothing via a sliding window to preserve continuity and reduce abrupt transitions between repaired sections and the original log data.
• Zone Based Repair: You can tailor repairs to geological and petrophysical intervals, which helps preserve local context

Well log data repair using the Auto Edit module in IP.

Benefits of Repairing Poor Quality Data Using the IP Auto Edit Module

Auto Edit is a powerful tool for enhancing your well log data quality in ways including:

• Improving Confidence: Build confidence in your data by applying a transparent and intelligent process that delivers clean and reliable well logs ready for further analysis and modelling.
• Maintaining Consistency: Auto Edit allows you to maintain a repeatable and systemic approach to repairing well log data between wells and projects to ensure a consistent data repair workflow.
• Increasing Efficiency: Spend less time manually editing curves by automating the cleaning process whilst maintaining transparency.
• Making Smarter Decisions: Clean and repaired logs are key to more accurate interpretations, better subsurface models and more reliable predictions leading to stronger and more reliable decisions.
• Streamlining Your Workflow: Auto Edit fits seamlessly into your interpretation workflow, bringing smart data repair to where you need it.
  
  

Better Logs, Better Decisions 

Poor quality well log data can have impact subsurface interpretations and petrophysical machine learning models, leading to key decisions being based upon potentially unreliable or incorrect data.

The Auto Edit module within IP provides a smart and systematic approach to repair and clean well log data using regression based algorithms with flexible and smart logic.

By improving data quality early in your petrophysical interpretations, Auto Edit can help you make more confident, consistent and effective decisions.

Further Reading

Sources: Banas, R., McDonald, A., and Perkins, T. J., 2021. Novel Methodology for Automation of Bad Well Log Data Identification and Repair. In SPWLA 62nd Annual Logging Symposium 2021.

https://www.spwla.org/SPWLA/Publications/Publication_Detail.aspx?iProductCode=SPWLA-2021-0070